A Novel Image Processing Filter Designed Using Discrete Fourier Invariant Signal

Abstract

In this paper, a new image processing filter is proposed. In order to construct this image smoothing filter, one dimensional discrete Fourier invariant signal generated by an iterative design principle based on gradient descent method is used. The filter shape and values in spatial and frequency domain is almost the same. The proposed filter can be used as a kernel matrix in image processing to perform blurring as well as high frequency noise suppression. Also it can be used as an optimal two dimensional window for spatial-frequency spectral analysis of images.